diff --git a/debug/misc/Makefile b/debug/misc/Makefile
index ccf89e4..2066b0d 100644
--- a/debug/misc/Makefile
+++ b/debug/misc/Makefile
@@ -18,6 +18,7 @@ endif
 
 obj-m := wei.o
 obj-m += uk_io.o
+obj-m += vfs_rw.o
 
 KERNEL_DIR ?= /lib/modules/$(shell uname -r)/build
 KERNEL_BUID_OUTPUT ?=$(KERNEL_DIR)
diff --git a/debug/misc/vfs_rw.c b/debug/misc/vfs_rw.c
new file mode 100644
index 0000000..252934d
--- /dev/null
+++ b/debug/misc/vfs_rw.c
@@ -0,0 +1,67 @@
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/fs.h>
+#include <linux/uaccess.h>
+
+/* read write buffer */
+static char wbuf[] = "this is a kernel read write test";
+static char rbuf[1024];
+
+#define FILE_NAME "/tmp/vfs_rw_test"
+
+/*
+ * 在内核中一般不容易生成用户空间的指针
+ * 或者不方便独立使用用户空间内存
+ * 而vfs_read,vfs_write的参数buffer都是用户空间指针
+ * 所以需要调用set_fs目的是为了让内核改变对内存地址检查的处理方式
+ * set_fs参数只有两个USER_DS和KERNEL_DS
+ * USER_DS代表用户空间,set_fs
+ * KERNEL_DS代表内核空间
+ */
+static int vfs_rw_init(void)
+{
+	struct file *fp;
+	mm_segment_t fs;
+	loff_t pos;
+
+	printk("%s, %d\n", __FUNCTION__, __LINE__);
+
+	/* open file */
+	fp = filp_open(FILE_NAME, O_RDWR | O_CREAT, 0644);
+	if (IS_ERR(fp))
+	{
+		printk("create file error\n");
+		return -1;
+	}
+
+	/* write file */
+
+	/* get current fs*/
+	fs = get_fs();
+
+	/* 即对内核空间地址检查并做变换 */
+	set_fs(KERNEL_DS);
+	pos = 0;
+	printk("=> %s\n", wbuf);
+	vfs_write(fp, wbuf, sizeof(wbuf), &pos);
+
+	/* read file */
+	pos = 0;
+	vfs_read(fp, rbuf, sizeof(wbuf), &pos);
+	printk("<= %s\n", rbuf);
+
+	/* close file */
+	filp_close(fp, NULL);
+	set_fs(fs);
+
+	return 0;
+}
+
+static void vfs_rw_exit(void)
+{
+	printk("%s, %d\n", __FUNCTION__, __LINE__);
+}
+
+module_init(vfs_rw_init);
+module_exit(vfs_rw_exit);
+MODULE_LICENSE("GPL");